simplify exec from init's subthread
Subsystems (2)
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH 1/1 simplify exec from init's subthread
[ On top of "exec: allow init to exec from any thread.",
filename exec-allow-init-to-exec-from-any-thread.patch ]
I think it is enough to take tasklist_lock for reading while changing child_reaper:
Reparenting needs write_lock(tasklist_lock)
Only one thread in a thread group can do exec()
sighand->siglock garantees that get_signal_to_deliver() will not see a stale value of child_reaper.
This means that we can change child_reaper earlier, without calling zap_other_threads() twice.
"child_reaper = current" is a NOOP when init does exec from main thread, we don't care.
fs/exec.c renamed from fs/exec.c~ +10 -12
@@ -616,6 +616,15 @@ static int de_thread(struct task_structkmem_cache_free (sighand_cachep, newsighand);return -EAGAIN ; } + +/* +* child_reaper ignores SIGKILL, change it now. +* Reparenting needs write_lock on tasklist_lock, +* so it is safe to do it under read_lock. +*/ +if (unlikely (current->group_leader == child_reaper)) + child_reaper = current; +zap_other_threads (current);read_unlock (&tasklist_lock);
@@ -660,23 +669,12 @@ static int de_thread(struct task_structstruct dentry *proc_dentry1, *proc_dentry2;unsigned long ptrace; - leader = current->group_leader; -/* -* If our leader is the child_reaper become -* the child_reaper and resend SIGKILL signal. -*/ -if (unlikely (leader == child_reaper)) { -write_lock (&tasklist_lock); - child_reaper = current; -zap_other_threads (current); -write_unlock (&tasklist_lock); - } -/* * Wait for the thread group leader to be a zombie. * It should already be zombie at this point, most * of the time. */ + leader = current->group_leader;while (leader->exit_state !=EXIT_ZOMBIE )yield ();
Quoted 10 lines · Oleg Nesterov
[ On top of "exec: allow init to exec from any thread.",
filename exec-allow-init-to-exec-from-any-thread.patch ]I think it is enough to take
tasklist_lockfor reading while changingchild_reaper:Reparenting needs
write_lock(tasklist_lock)Only one thread in a thread group can do
exec()
Quoted 8 lines
sighand->siglockgarantees thatget_signal_to_deliver()will not see a stale value ofchild_reaper.This means that we can change
child_reaperearlier, without callingzap_other_threads()twice."
child_reaper= current" is a NOOP when init does exec from main thread, we don't care.
Looks good. I fat fingered my locking anyway, and this looks like a good fix, as well as the more obvious place, to do this.